On Mon, Apr 08, 2002 at 09:32:37PM +0900, Dan Kogai wrote: > On Monday, April 8, 2002, at 08:41 , Graham Barr wrote: > > This is a bug in HTTP::Date which the latest Time::Local::localtime show > > and previous version did not. I thought a fixed HTTP::Date had been > > released. > > > > Graham. > > I would like to make this sure; Does this imply the previous > Time::Local was buggy for returning fractional value for fractional > second?
Yes. This is just one of the many bugs in it. > Well, so far as I see localtime() always returns an integral value (uses > struct tm, which values of member are integral for most, if not all, > cases). Therefore the following equivalence > > $whenver == timelocal(localtime($whenever)) > > will not stand unless timelocal() returns an integral value. But > HTTP::Date took advantage of timelocal()'s undocumented feature, only to > fine this backdoor was closed. > > Did I get the picture correctly? Yes. Graham.
